Publications

6 Results
Skip to search filters

Training neural hardware with noisy components

Proceedings of the International Joint Conference on Neural Networks

Rothganger, Fredrick R.; Evans, Brian R.; Aimone, James B.; DeBenedictis, Erik

Some next generation computing devices may consist of resistive memory arranged as a crossbar. Currently, the dominant approach is to use crossbars as the weight matrix of a neural network, and to use learning algorithms that require small incremental weight updates, such as gradient descent (for example Backpropagation). Using real-world measurements, we demonstrate that resistive memory devices are unlikely to support such learning methods. As an alternative, we offer a random search algorithm tailored to the measured characteristics of our devices.

More Details

Development characterization and modeling of a TaOx ReRAM for a neuromorphic accelerator

Marinella, Matthew J.; Mickel, Patrick R.; Lohn, Andrew L.; Hughart, David R.; Bondi, Robert J.; Mamaluy, Denis M.; Hjalmarson, Harold P.; Stevens, James E.; Decker, Seth D.; Apodaca, Roger A.; Evans, Brian R.; Aimone, James B.; Rothganger, Fredrick R.; James, Conrad D.; DeBenedictis, Erik

This report discusses aspects of neuromorphic computing and how it is used to model microsystems.

More Details
6 Results
6 Results